INDEX

Explanations

separate list items

The neuron activates on discourse-level adverbs and transition markers (e.g. actually, though, however, then, course, too).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

-1.92

了一

-1.75

Karakteristik

-1.74

鉆

-1.73

 дает

-1.73

 allergy

-1.63

 دوباره

-1.59

 appetizer

-1.57

POSITIVE LOGITS

1.92

@[+][

1.84

上で

1.84

 другим

1.77

 elucid

1.76

mathrm

1.74

क्टर

1.73

eynman

1.71

viene

1.67

aarrggbb

1.59

Activations Density 0.067%